We Built it, We've measured it, HELP US TWEAK IT - acoustics of the BLACK CAT - Page 3

So, how's about you try measuring the other speakers (outside if possible)?

You don't need to go outside for that...
Put the microphone right up to the tweeter...not touching, but almost. Play full scale pink noise (internal test tones in pre-pros are not full scale, many test DVD tracks are not full scale unless specifically so noted).

If you get the same result, try another speaker ... one you know works. That will begin to point you to tweeter or mic problems.

Oh, thank goodness, it would've been quite a hassle to take the speakers outside...
I thought I was going to have to get a really long speaker cable from the hardware store :ugh:

And premium Monster cable is pretty expensive

Wouldn't it still be useful to have the speaker free space frequency response?

If the microphone is close enough to the speakers you are, for most intents and purposes, eliminating room effects.

The DE article HopefulFred posted on the 23rd says to have the mic "1' meter" [sic] away - we assumed it was "1 meter" but, based on this discussion, maybe we should've assumed "1'" (1 foot). Actually, now it seems that 1" (1 inch) might be even more preferred...is that right?

Big is correct.

So that means the speaker still needs to go outside to measure the overall FR?

@BritInVA - Monster cable, smonster cable The 12ga solid copper cable from HD works great!

You can schlep your stuff outside (if you do this, understand you'll have to hang that speaker more than 3.5' above the ground). You could also put the speaker under test in the middle of your room. Here's a math exercise for you:

If your microphone is 1 meter from the speaker, how long will it take for the sound to travel from the speaker to the microphone?

What is the nearest surface to the speaker (likely the floor). What is the distance sound will travel from the speaker to the floor to the microphone? What is the amount of time for sound to travel that distance? Set the measurement window to be less than the time from speaker to floor to microphone.

There is one "gotcha" here. Processor delay! Be sure to use "direct mode" (neutralize all speaker delays, tone controls and other nonsense) on your receiver, run an ETG. What is the elapsed time between when your gear sent the first "ping" or "click" and the arrival of that sound at the microphone. The difference between the actual elapsed time and what the elapsed time should be based upon the speed of sound, is the processor delay ... so you need to accommodate for that in your measurement window timing.

*AND*, of course, what you did before this measurement is measure and plot the noise floor in the room.

He'll need to know a few more variables........ie. atmospheric pressure, humidity levels, temperature etc........

How about just starting at 1m/(343 m/sec) = 0.0029 secs

All joking aside, wouldn't they need to worry about wind affecting the results?

Anything cooking around here?

In the immortal words of the Boss in Cool Hand Luke, "What we have here is a failure to communicate..."

I was on business travel and then a week at the beach, so not not time to do stuff. That said, I now have REW completely setup and playing sweeps through speakers. I only did calibration tests so far, so no measurements to report just yet.

I was thinking more about what low-hanging options remain: replace some linacoustic with batting, put some scrim/paper/plastic on top of Linacoustic, add a 2nd subwoofer, etc. The thing I'm trying to understand is how these changes can be measured and then see how they go. For instance, I think the answers are better ETC, better mid/high-range decay, and flatter/louder bass response.

One thing that I've found interesting is that some AT plans I've seen or discussed with board members about didn't specify things like if there would be chair/crown molding or what material the chairs are made of. I zoomed in on these points in particular because it seems that I have a lot of molding/furring (about 33% of the wall) and because my seats are fabric (some with temperpedic cushion) - the fabric isn't designed to be AT, but it's hard to imagine it being as reflective as leather either, at least to some frequencies...

After some time measuring and staring at graphs, I'm finally ready to post an update (yeah!)

First, I must confess that it's been difficult to stay focused on this project, with the room being fully functional and what not. I've watched and number of great movies with family and friends already, everyone seems to love the space as it is, but I push on.

Cool, so below are my results for the Nyal Mellor criteria Helpful Fred posted:


Here's the SPL of my room with nothing playing (I disconnected the output cable). From looking around online, it seems that this might be RC35. To be honest, I'm a bit surprised that it's not lower, with the decoupled walls and ceiling. What are other people getting? As for the spike around 60Hz, I think that the the heat pump running - the unit is right outside this room in the basement...




The two pictures below are for all three front speakers - first with no smoothing and then with 1ms smoothing. I can't say for sure, but it seems OK to me...???







I usually look at the waterfall graphs, but since Nyal's article uses spectrograms, I'll show them too. Note, for each speaker, both the sub and the speaker are playing.

First the waterfalls (spectrograms next):




















And now the spectrograms (waterfalls before):

















This Bass Integration Guide by Paul Spencer suggests that there should be a 20dB decay after 150ms. I couldn't figure out how to get GIMP to plot the -20dB line, but I think that it's not too hard to eyeball...


















I'm not sure if I generated these graphs correctly, but I think they might indicate an overly short decay for the midrange (250Hz - 4000Hz) - right?


















OK, so my speakers seem to be more than +-3dB, it's more like +-5db





For the 1/24th octave, 108Hz goes from 80-105dB, 169hz goes 84-112dB, and 183Hz goes 78-107dB

For the 1/3rd octave, 173Hz goes 90-104dB and 194Hz goes 87-103dB









My room is roughly 315 sq ft and 2700 cu ft

Calculating if the room has common divisors is difficult to do since the walls/ceiling aren't perfectly straight/level - no single measurement can be used. Also, my room has a bump-out behind the screen, where it's under a a bay window, so the length of the room is an extra 2.5 feet in the area.

Yes, the room is DD+GG on top of decoupled walls and ceiling.





OK, so this was a much longer post than I expected. I actually kind of knew it would be, as I spent some time yesterday thinking about all the pics that would have to be posted

BTW, I tried to test the tweeter drop-off issue/concern from before. To be honest, I'm not sure what the concern is, other than it looks like tweeter has little output from beyond 10kHz. I'm not sure if the below graph helps, but it seems that *all* the speakers are dropping off in that range. Now, admittedly, this is a farfield measurement (as are all the others), so maybe this indicates too much high-freq absorption - would putting scrim/plastic/paper over the linacoustic help?




Anybody have any advice for how to make this room sound better?

I have a similar thread going for my room and I have the same questions about the room sweep. You can see I did 3 diff measures, but getting the laptop & myself out of the room was a big difference.
I'm not sure what is correct. FWIW- no isolation in use.
http://www.avsforum.com/t/1415741/speaker-measurement-room-measurement-treatment/30#post_22179344

Also, did you apply any smoothing at all for your room sweep?

I hope you get lots of responses on this post. I'll think for a while before posting much.

I think you should feel okay about your RC data. Dennis has suggested that "prosumer" gear is inadequate for good noise floor measurements - well, at least I remember him saying that doing it properly requires thousands of dollars worth of equipment (mics especially, IIRC). So the fact that your measurement is in the ballpark, I take for a good sign. Yeah, that 60Hz peak looks bad on paper... to me anyway. But there's nothing reasonable to do about it, right? Are you sure it wasn't ground loop hum?

Your RC/NC/NR is a curve fit exercise. If your equipment does not specifically do that for you, you'll have to do it yourself. (The "curve fit" exercise in the most basic of terms compensates for our hearing sensitivity by frequency). The 60 Hz bump could be an issue with your equipment ... if you think it is your outside unit, just turn the unit off and remeasure. A noise floor of 30 to 35dBSPLA in a residence is pretty common. You need to start looking for potential weak links in your isolation/construction...doors, windows are weak links, noise from stuff you brought into the room, potential flanking paths, etc.

Tweeter drop off. This could be any number of issues. It is suspect that (1) all the curves have the same shape ... a gradual roll off from 2kHz, bumping back up at 8kHz, peaking at 10kHz, then falling off the cliff; (2) side right, side left have a different SPL than the other speakers; and, (3) it appears the L/C/R track pretty closely to each other. The missing evidence is the near field measurements (ie, a graph showing the near field overlaid on the above data...one speaker at a time). (The nearfield will can eliminate room effects.) Since high frequencies will roll off over distance at a higher rate than lower frequencies, you'd want your nearfield to reflect a rise in SPL from about 10kHz to 20kHz to compensate for listening distance. Other than poor crossover design, or blown tweeters the more common issue is the inability of the measurement kit to accurately measure above 10kHz. None-the-less, assuming a lack of measurement accuracy from 10kHz to 20kHz, the concern is the rolloff from 2k to 8kHz...that should not be there. To determine if that is a speaker issue or room treatment effect, you have to overlay the nearfield results.

I just read the whole thread and subscribed to it. It's great to see someone else crazy enough to try the empirical method

Your thread is currently focused on 1-meter tests and how to remove processor delay. In contrast, I jumped straight to testing the primary listening position (farfield). Your analysis is looking for 1st-reflection points in an untamed room, undoubtedly to help guide towards knowing which treatments to place where, whereas my room is already loaded with treatments, so I'm more interested in the broad-picture (i.e. how it matches up to the criteria in Nyal's report).

FWIW, I am still interested in tracking down nasty reflections of course, in fact I found a surprising one already (my side surround speaker bouncing off the soffit (not posted, since the ETC was only for the front speakers). I'm not sure important this is, though, since it's just a surround speaker and putting treatment in that location would be difficult to make look nice - essentially, this might be a compromise I'm willing to live with

I don't understand, I thought smoothing was a post-capture filter - in REW's "Graph" menu option, you can apply smoothing for different octaves. From your thread, I saw that dragonfly recommended no smoothing too, for when posting data files. so there must be a capture time thing option. AFAIK, I'm using REW's default smoothing, whatever it is...

Cool beans. If you or anyone else is interested, I posted the REW mdat files here: http://softnami.com/theater/2012-07-09/farfield-measurements.tgz (17.3 MB)


I forgot to post earlier - I'm using the Tascam US-144 sound card and the Dayton Audio EMM-6 microphone. In REW, a calibrated the sound card using a loopback and I fed it the calibration file for the mic from Dayton's site. I understand that it's amateur/hobbyist gear, but hopefully sufficient for my goal...


As Dennis suggests, it could've been the outside HVAC unit running. I'll measure again after turning off all the units...

I was asking about the room sweep in particular, because my sweep was not as smooth as yours. It looked just like a speaker sweep without smoothing and I had to smooth it out to get something semi understandable. I was just trying to bounce info off your experience.

I've resolved my delay issues, so I'm onto the next step. I'm planning to do it all, but I prefer to learn it in steps. I'm watching this thread as well.

Hi Dennis, I'll remeasure with the outside units off. Interestingly, I notice the BIG's "quite room" measurement (post #8) also showed a spike at 60Hz, along with the others that my equipment didn't pick up. I don't think this means anything yet, at least not until I measure again with the outside units turned off, we'll see...

I'm curious about your saying 30-35dB SPL is normal for residential space - does that include dedicated theaters built by DIYers too? My space is DD+GG on decoupled walls/ceiling, acoustical chalk in every crack that ever presented itself, putty pads on every outlet/switch box, two solid-core communicating doors (one facing theater is sealed on all four edges), and the theater's recirculating fan was off. I guess I would've thought that all that might get it to RC25, but you're saying that it's almost as if I did nothing at all - how can that be right?


I'll definitely do the nearfield measurements next. Do you think it's better to do it outside on a calm day or to bring each speaker to the center of the room?

BTW, BIG's 1-meter "on axis" measurement looked pretty flat (post #41), but both his "off axis" and farfield measurements showed a midrange roll off - the weird thing is, his graph show it bottoming-out around 1200Hz, whereas mine show it bottoming-out around 8kHz. Again, wildly different equipment and software...hopefully my nearfield measurements come out good enough to eliminate concern for my mic/soundcard...

I cannot tell you the number of times a DIYer or local contractor thought they had done all the right things right and didn't. In fairness, the are bunch that the right things were done right. It is possible you're measuring the noise floor of measurement kit itself.

Is there a good way to know this? What would you think is the quietest place you could measure? Maybe the inside of a refrigerator that was unplugged (while the a/c is also turned off - with the mic wrapped in pillows)?

Googling a little bit, I found a couple folks saying it had a floor around 33dBA, which matches my results. The specs for the mic don't seem to tell:

How about a cooler inside a cooler inside a cooler inside my cool theater? I can try that, but the hard part will be feeding the cable though the lid. Maybe I'll get lucky and find that some of my coolers have drains that the cable can feed through...

I'd like to resolve the 60Hz spike issue, but I don't think the mic being able to measure the room's true floor matters much, as the test tones are a solid 50dB over the floor. That said, it would be nice to know how quiet the room is

Measuring outside eliminates all reflections except those from the ground. I'm sure there is a difference between sound bouncing off grass vs carpeted-concrete, but I've never seen the absorption/diffusion characteristics described before. Is the 3.5' number somehow special for outside measurements?


Sound would travel 1m in 1/343 sec == 3ms. Assuming the floor is 3.5ft (1.07m) off the ground, the bounce would arrive in 6.9ms. Ignoring processor delays for now, subtracting the two values sets the window to be 3.9ms, which would allow accurately measurement of frequencies down to 256Hz.

In order to negate the floor-bounce, some people lay insulation on the floor between the speaker and mic. It seems that this might only allow not needing to set the window/gate for higher frequencies, down to whatever the insulation can absorb. This goes back to measuring outside, is grass-covered-earth a great absorber by any chance?